sc_reconstruction.metrics.metric_cellcycle#

sc_reconstruction.metrics.metric_cellcycle(adata_true, adata_pred, *, s_genes, g2m_genes, min_cells=20)[source]#

Cell-cycle phase agreement between true and predicted expression.

Cell-cycle scoring follows [Tirosh et al., 2016] (the S- and G2M-phase gene lists shipped with most single-cell pipelines). Wraps CellCycleCalculator.cell_cycle_labeling_similarity().

Parameters:
  • s_genes (Sequence[str]) – Gene symbols (matching adata_*.var_names) defining S and G2M phases. Pass the standard [Tirosh et al., 2016] lists or a custom split.

  • g2m_genes (Sequence[str]) – Gene symbols (matching adata_*.var_names) defining S and G2M phases. Pass the standard [Tirosh et al., 2016] lists or a custom split.

  • min_cells (int) – Minimum number of cells per gene for inclusion.

  • adata_true (anndata.AnnData)

  • adata_pred (anndata.AnnData)

Return type:

dict[str, float]